How SnapMirror works with the dump command

How to back up data in the mirror

If you want to back up data from a mirror, the data must be in an existing snapshot, not the active file system, on the mirror. This is because before the dump command writes data from an active file system to tape, the filer must create a snapshot in the volume containing the data. The filer cannot create snapshots on the mirror, which is a read-only volume, to be used by the dump command.

You can back up any snapshot displayed by the snap list command on the mirror. You can also create a snapshot on the source volume, replicate the snapshot to the mirror, and use the dump command to back up this snapshot from the mirror to tape.

Effect of the dump command on the mirror update schedule

Running the dump command on a mirror might change the mirror update schedule in the following ways:

If the mirror is scheduled to be updated when a dump command is in progress on the mirror, the filer delays replicating data from the source volume until the dump command is finished.
If you start the dump command when the filer is replicating data to the mirror, the filer stops the replication and does not update the mirror. It restarts replication after the dump command is finished.

Delaying the mirror update is necessary because when the dump command backs up data from a snapshot, the snapshot must exist until the command is completed. During a mirror update, the filer deletes existing snapshots on the mirror and copies new ones from the source. This process causes the snapshot being backed up to disappear and the dump command to terminate.

Part No. 210-02334
File: mirror5.htm
Last updated: 04/26/99 08:11:54